BNL, Upton, Long Island, New York, USA
With the aid of machine learning techniques, the genetic algorithm has been enhanced and applied to the multi-objective optimization problem presented by the dynamic aperture of the NSLS-II Ring. During the evolution employed by the genetic algorithm, the population is classified into different clusters. The clusters with top average fitness are given elite status. Intervention is implemented by repopulating some potentially competitive candidates based on the accumulated data. These candidates replace randomly selected candidates among the original data pool. The average fitness of the population is improved while diversity is not lost. The quality of the population increases and produces more competitive descendants accelerating the evolution process significantly. When identifying the distribution of optimal candidates, they appear to be located in isolated islands within the search space. Some of these optimal candidates have been experimentally confirmed at the NSLS-II storage ring. The machine learning techniques that exploit the genetic algorithm can also be used in other population-based optimization problems such as particle swarm algorithm.
TRIUMF, Vancouver, Canada
UIS, Bucaramanga, Colombia
UMAG, Santa Marta, Colombia
The results of the relativistic full electromagnetic Particle-in-cell (PIC) simulation of a bunched electrons beam accelerated in a TE113 cylindrical cavity in the presence of a static inhomogeneous magnetic field are presented. This type of acceleration is known as Spatial AutoResonance Acceleration (SARA)*. The magnetic field profile is such that it keeps the electrons beam in the acceleration regime along their trajectories. Numerical experiments of bunched electrons beam with the concentrations in the range 108 -109 cm-3 in a linear TE113 cylindrical microwave field of a frequency of 2.45GHz and an amplitude of 15kV /cm show that it is possible accelerate the bunched electrons up to energies of 250keV without serious defocalization effect. A comparison between the data obtained from the full electromagnetic PIC simulations and the results derived from the relativistic Newton-Lorentz equation in a single particle approximation is carried out. This acceleration scheme can be used to produce hard x-ray**.
* Dugar-Zhabon, V. D., & Orozco, E. A. (2009).Physical Review Special Topics-Accelerators and Beams, 12(4), 041301.
** Dugar-Zhabon, V. D., & Orozco, E. A. (2017). (USA Patent: 9,666, 403 )
BNL, Upton, Long Island, New York, USA
Electron polarization in a storage ring is subject to two very long term effects: Sokolov-Ternov polarization and depolarization by diffusion. This leads to an equilibrium state over a very long time scale, and, simulation-wise, is highly CPU-time and -memory consuming. Simulations aimed at determining optimal ring storage energy in an electron-ion collider in this context, are always based on tracking bunches with thousands of particles, and in addition for short time scales in comparison, due to HPC limitations. Based on considerations of ergodicity of electron bunch dynamics in the presence of synchrotron radiation, and on the very slow depolarization aimed at in a collider, tracking a single particle instead is investigated, here. This saves a factor of more than 2 orders of magnitudes in the parameter CPU-time*Memory-allocation, it allows much longer tracking and thus improved accuracy on the evaluation of polarization and time constants. The concept is illustrated with polarization lifetime and equilibrium polarization simulations at the eRHIC electron-ion collider.
BNL, Upton, Long Island, New York, USA
We present our experiences in optimizing the proposed Rapid Cycling Synchrotron (RCS) injector for the eRHIC Storage ring and the RHIC 2017 lattice. We have develop python code to drive lattice calculations in MADX which are then used to calculate spin resonances using the DEPOL algorithm. This approach has been used to minimize intrinsic spin resonances during the RCS acceleration cycle while controlling lattice parameters such as dispersion and beta functions. This approach has also been used to construct localized imperfection bumps using a spin response matrix and SVD. This approach has also been used to reduce interfering intrinsic spin resonances during the RHIC acceleration ramp.
UMD, College Park, Maryland, USA
Invited talk: Design of accelerator lattices with nonlinear optics to suppress transverse resonances is a novel approach and may be crucial for enabling low-loss high-intensity beam transport. Large amplitude-dependent tune spreads, driven by nonlinear field inserts, damp resonant response to driving terms. This presentation will focus on simulations of the UMER lattice operated as a quasi-integrable system (1 invariant of transverse motion) with a single strong octupole insert. We will discuss the evolution of simulation models, including the observation of losses associated with the original operating point near a fourth-order resonance. Other operating points farther from this resonance are considered and shown to be more promising.